Please choose your delivery country and your customer group
This paper treats the influence of a stress gradient of smooth and notched components on the fatigue life. Due to the difficulties related to crack growth analyses in real components, fatigue lives are often predicted by means of the maximum stress at the component surface. In many instances, this approach may yield satisfactory solutions, especially since life estimates will always be on the safe side. However, the maximum stress approach may sometimes lead to overconservative estimates. This paper has quantified the degree of conservatism and identified features influencing the normalised lifetime by performing crack growth analyses both on the real and the smooth specimens. Two cases were considered: a semi-infinite smooth specimen subjected to a linearly decreasing stress and a semi-infinite notched specimen subjected to a uniform remote stress. The degree of conservatism of the maximum stress approach was quantified by normalising the lifetime calculated for the actual specimen with respect to the lifetime of a smooth homogeneously stressed specimen subjected to the same maximum stress. The normalised lifetimes were presented for various initial crack depths as a function of the relative stress gradient and notch curvature. Generally, there are two main features influencing the degree of conservatism: (i) the initial crack depth, ai, and (ii) the stress field in the plane of the crack, quantified by the stress gradient or the notch curvature. Deep initial crack depths, ai, and high relative stress gradients lead to overconservative estimates, especially for edge through-cracks. Using the maximum stress approach gives the best results for semi-elliptical cracks with shallow initial cracks ai, where good lifetime estimates are achieved even for relatively high stress gradients.